基于混合效应的湖南栎类与杉木单木断面积生长模型研究

Study on Growth Model of Individual Tree Basal Area of Quercus sp.and Cunninghamia lanceolata Forest Based on Mixed Effect in Hunan Province

构建栎类(Quercus sp.)杉木(Cunninghamia lanceolata)混交林单木断面积生长模型,为该类型林分的生产经营提供参考。以湖南省31个样地共1 646株栎类与1 154株杉木为研究数据,7个常见模型作为栎类、杉木的单木基础模型,基于样地随机效应构建栎类、杉木的单木断面积混合效应模型。结果表明,栎类断面积的最优基础模型为Gompertz,杉木断面积的最优基础模型为Logistic;栎类断面积混合效应模型相比于基础模型,决定系数R^2提高了0.053,SSE与RRMSE均有所降低;杉木断面积混合效应模型相比于基础模型,决定系数R^2增加了0.035,SSE与RRMSE均有所降低;栎类断面积的成熟年龄在50年以后,杉木断面积的生长速率在30年左右逐步放缓。文章在栎类、杉木单木模型的基础上进一步加入样地随机效应,模型拟合精度明显提高,说明栎类、杉木的断面积生长都受到年龄与样地的综合影响。

The growth model of basal area of individual tree in mixed forest of Quercus sp.and Cunninghamia lanceolata was constructed to provide reference for the production and management of this type of forest.Based on the data of 1646 Quercus sp.and 1154 C.lanceolata in 31 plots in Hunan province,seven common models were used as the individual tree foundation models of Quercus sp.and C.lanceolata,and the mixed effect models of single wood basal area of Quercus sp.and C.lanceolata were constructed based on the sample random effect.The results showed that the optimal basic model of basal area of Quercus sp.was Gompertz,and the optimal basic model of basal area of C.lanceolata was Logistic.Compared with the basic model,the determination coefficient R2 of basal area mixed effect model of Quercus sp.increased by 0.053,while SSE and RRMSE decreased.Compared with the basic model,the determination coefficient R2 of basal area mixed effect model of C.lanceolata increased by 0.035,while SSE and RRMSE decreased.The mature age of Quercus sp.basal area was above 50 years,and growth rate of C.lanceolata slowed down gradually in about 30 years.After further considering the random effect of sample plots,the fitting accuracy of Quercus sp.and C.lanceolata individual models were significantly improved,it indicated that the basal area growth of Quercus sp.and C.lanceolata was influenced by age and plot.